News for hackers. Stuff that matters.
Re:Dialects of Lisp by brudgers on Friday February 22, 2013 @12:33PM (#5265411)
If I were to bet, I'd place my money on Clojure as the 100 year language. It is as unbound by hardware in regard to cross platform development as any other likely candidate, it has good access to libraries meeting contemporary expectations, and is likely to continue to have access to new libraries written to meet future expectations - all due to its interaction with the JVM (and .NET).

Of course I am adopting PG's premise that the 100 year language is a LISP while ignoring his premise that it is not a bit kludged up by practical considerations (i.e. the ugliness of its dependence on the JVM).

(/edit) Clojure also has an evangelist in Rich Hickey whom people seem to respect independently of Clojure.

Re:Dialects of Lisp by pavelludiq on Friday February 22, 2013 @01:08PM (#5265606)
Common Lisp is 29 years old(if we count from the publication of CLTL1) or 19 years old(if we count from the publication of the ANSI standard), clojure is 5-6 years old. Scheme is even older. I love clojure, but it has some mighty big shoes to fill, CL was build for longevity from the beginning. Also, both scheme and CL run on the JVM. CL also survived the death of many operating systems, architectures and platforms in those 29 years, clojure is yet to survive one.
Re:Dialects of Lisp by brudgers on Friday February 22, 2013 @01:37PM (#5265787)
To me, Common Lisp appears to suffer from bad timing. It was designed on the cusp of paradigm shifts from timesharing to networking; from text interfaces to graphic interfaces; and from computers being rare to being ubiquitous. Hitting it's peak in popular consciousness in the midst of the Black Friday and S&L recession didn't help either.

My view is premised upon the idea that a major deficit of Common Lisp is a poor access to contemporary independent user interfaces. The solutions that exist are not competitive with the leverage Clojure enjoys from JVM (and .NET).

Re:Dialects of Lisp by pavelludiq on Friday February 22, 2013 @02:15PM (#5266081)
You are somewhat correct, but one of the reasons I love common lisp is exactly because it sort of overcame this "bad timing", which is very unforgiving to lesser languages. The fact that today I can write lisp apps in a language whose primary platform(the lisp machine) died 20 years ago, and still enjoy a relatively active ecosystem is quite impressive.

Do you think clojure has the potential to be useful in 20 years if the JVM dies today hypothetically? I don't know, I think yes, because it has a bunch of good ideas, but it isn't at all clear. Take JS as an opposite example, if the web is to disappear, JS will disappear as well, no matter how many other platforms it is ported to, it just isn't good enough to stand on its own.

I could be talking out of my ass here, but my point was that CL has a good track record for survival, so it has the potential for a 100 year language, while clojure is yet to prove itself for such a task. These things are extremely hard to predict, tech history has a habit of throwing such wild-cards as personal computing and the web. We'll see.

Re:Dialects of Lisp by lispm on Saturday February 23, 2013 @05:01AM (#5269184)
The primary platform of CL did not die. CL was ALWAYS developed on multiple platforms. From day -1 there were implementations on non Lisp Machine platforms. People actually ran companies to do so: Lucid (Unix), Franz (Unix, PC), Coral (Mac), Goldworks (PC), Procyon (PC, Mac), Harlequin (Unix), ... The Carnegie Mellon University developed CMUCL for Unix. Japanese developers KCL (Unix, ...), CLISP from Germany, ...
Re:Dialects of Lisp by mark_l_watson on Friday February 22, 2013 @02:27PM (#5266181)
You are correct, I think, on the bad timing. Also factor in that Common Lisp used to be expensive to use and deploy. In the early 1980s, I 'went cheap' and bought a Xerox 1108 (for $25K, which was a lot of money in those days). Deployment costs? Buy a 1108 for each customer. Yuck! I used my 1108 for prototyping and demos (and made money selling an AI product written for it), but no deployed systems.

Today with free SBCL and Clozure (among many others), deployment is free - what I would have given in the 1980s for something like SBCL!

Re:Dialects of Lisp by lispm on Friday February 22, 2013 @05:03PM (#5267171)
We deployed Common Lisp on Unix for nothing using KCL/AKCL/GCL. AKCL appeared in the mid 80s.

In the 80s SBCL was called CMUCL. Free.

Re:Dialects of Lisp by mark_l_watson on Friday February 22, 2013 @06:50PM (#5267720)
That is true. I don't remember why I didn't use CMUCL. We were mostly a Dec VAX shop; perhaps VAX was not supported?
Re:Dialects of Lisp by lispm on Saturday February 23, 2013 @05:02AM (#5269190)
If you were using DEC Common Lisp, you might have used bits of CMUCL.
Re:Dialects of Lisp by alberich on Friday February 22, 2013 @01:45PM (#5265843)
My bet is that when Clojure gets to be 29 years old it will, also, suffer from several problems. I don't believe it is possible to design a future proof programming language. Who knows what will come next and how it will affect programming tasks?
Re:Dialects of Lisp by brudgers on Friday February 22, 2013 @02:03PM (#5265973)
To clarify, I am making a reference to this PG essay:

http://www.paulgraham.com/hundred.html

Re:Dialects of Lisp by jdale27 on Friday February 22, 2013 @09:26PM (#5268260)
Hitting it's peak in popular consciousness in the midst of the Black Friday and S&L recession didn't help either.

Is there any real connection here with CL? Did these events somehow precipitate or deepen the AI winter?

Re:Dialects of Lisp by mark_l_watson on Friday February 22, 2013 @02:37PM (#5266235)
Well, Rich used to use Common Lisp (I know because I used to use his very neat LispWorks to Java bridge stuff).

Clojure really does hit the sweet spot combining a concise agile language with good practical decisions, similar I think to some of the decisions that went into the design of Scala (like a generalized seq API for lots of data structures).

Except for some personal serious side projects, I am now retired, and I chose Clojure over other good possibilities (e.g., Common Lisp, Racket, Gambit-C, Scala, Smalltalk, or Ruby) because it is a practical language and still lots fun to code in.

Now that I don't have to worry about staying up to speed on all of the languages my customers used to like using, I find that just using a single language really saves me a lot of time. Some advice for people still working: perhaps choosing a single practical and extensible language like Scala or Clojure is a good strategy; I used to like being a polyglot programmer but that requires a lot of overhead!

Re:Dialects of Lisp by pmelendez on Friday February 22, 2013 @12:52PM (#5265508)
I wouldn't put my money on the JVM platform. It's funny but Clojure's bigest advantage is also its biggest disadvantage.
Re:Dialects of Lisp by brudgers on Friday February 22, 2013 @01:08PM (#5265602)
Given its adoption by enterprise, the JVM at worst has a COBOL future. At best, it's future is worse-is-better.

http://www.dreamsongs.com/WorseIsBetter.html

Re:Dialects of Lisp by Arelius on Friday February 22, 2013 @03:52PM (#5266704)
> At best, it's future is worse-is-better

If we're taking the worse-is-better angle My suspicion is that clojurescript, which compiles to javascript, is a much more likely candidate to continue based on worse-is-better.

Re:Dialects of Lisp by asynchrony on Friday February 22, 2013 @03:42PM (#5266640)
Clojure is so closely tied to the JVM that it's hard to imagine it being the "100 year language" unless the JVM turns out to be the "100 year platform".
Re:Dialects of Lisp by ldh on Friday February 22, 2013 @05:03PM (#5267172)
The original Clojure implementation was built on the JVM, but it also targets the CLR and JavaScript. I don't think it's irrevocably tied to the JVM.
Re:Dialects of Lisp by asynchrony on Friday February 22, 2013 @05:08PM (#5267198)
Some people would argue that once you remove the JVM there isn't much there (that can't be found elsewhere).
Re:Dialects of Lisp by dguaraglia on Saturday February 23, 2013 @03:00AM (#5268944)
Exactly! The reason why languages like Clojure and Scala are becoming popular is not only their cool features (although in that sense Scala has an advantage over Clojure which is just 'another Lisp') but because of the ready availability of libraries to do pretty much any thing you need to do. Take that away from Clojure, and you have something that can barely compete with CL. Take it away from Scala and you end up with another academically interesting, yet obscure language (like Haskell.)
Re:Dialects of Lisp by ema on Friday February 22, 2013 @05:29PM (#5267327)
Clojure is currently my favorite language, partly for the pragmatic choices it made, but these pragmatic choices will limit its continued use.

Hygienic macros often prevent one from shooting oneself into the foot, but they prevent also some truly elegant code. Having lists and vectors with a common interface is cool, but that the semantics of conj is different for both is annoying. Easy java interop is worth a lot, but it prevents call/cc.

It's probable that i will gladly use Clojure for the next 10, maybe even 20 years, but it won't be the programming language my grandchildren will ask me to shut up about.

Re:Dialects of Lisp by asynchrony on Friday February 22, 2013 @10:42PM (#5268449)
Hygienic macros prevent accidental symbol capture. They don't prevent intentional symbol capture. You may be thinking of the limitations of pattern based macros, but many Scheme systems provide hygienic procedural macros.
Re:Dialects of Lisp by Tyr42 on Friday February 22, 2013 @08:08PM (#5268011)
I am curious, what do you think hygienic macros break? I'm a fan of Racket, and I haven't yet found anything that it's macros can't do.
Re:Dialects of Lisp by nbouscal on Saturday February 23, 2013 @02:10AM (#5268860)
I'm still too new to Lisp to answer this question, but I have seen the book Let Over Lambda recommended as having answers to that and similar questions.
Re:Dialects of Lisp by crististm on Friday February 22, 2013 @12:41PM (#5265452)
Not bounded by hw but still bounded by its VM implementation.
Re:Dialects of Lisp by robdor on Friday February 22, 2013 @12:47PM (#5265483)
With Clojure being available for the jvm, the clr through ClojureCLR and javascript through ClojureScript I don't believe that this will be much of an issue.
Re:Dialects of Lisp by asynchrony on Friday February 22, 2013 @03:47PM (#5266681)
Except that ClojureScript is dependent on JVM Clojure, and if ClojureCLR were to take advantage of the CLR's capabilities (TCO and stack allocation, for example) it also wouldn't be Clojure.

So we're really talking about dialects of Clojure (and perhaps might as well be talking about dialects of Lisp in general).

Re:Dialects of Lisp by swannodette on Friday February 22, 2013 @04:22PM (#5266926)
We're not that far from officially having a boostrappable ClojureScript compiler. And some people have already done it - http://github.com/kanaka/clojurescript
Re:Dialects of Lisp by billsix on Friday February 22, 2013 @05:10PM (#5267215)
>It is as unbound by hardware in regard to cross platform development as any other likely candidate

Gambit c runs on more hardware